san francisco - onsite 4 days per week
Jr -0 mid level candidates
About The Role
We’re assembling a world-class research group focused on making that vision a reality. If you're excited by open-ended technical challenges and want your work to directly shape a product used at scale, read on.
What You’ll Do
As a foundational member of our research team, you'll work across the stack—from algorithm design to deployment—to unlock new capabilities in autonomous agents.
You will:
-
Invent and prototype new approaches in planning, memory, multi-agent systems, and learning from human feedback.
-
Deploy at scale, with real users and datasets, running experiments across large compute clusters and iterating rapidly.
-
Design practical systems, working closely with engineering and product teams to integrate research into production-ready tools.
-
Define best practices around model evaluation, safety, and alignment, helping establish a research culture focused on impact and rigor.
-
Mentor and collaborate, working with engineers, interns, and collaborators from different backgrounds to maximize team output.
What We’re Looking For
-
A graduate degree (or equivalent experience) in machine learning, reinforcement learning, NLP, or related fields.
-
A track record of high-quality research—through papers, open-source work, or real-world deployments.
-
Proficiency in frameworks like PyTorch or JAX, and hands-on experience training large models on distributed infrastructure.
-
A willingness to dig into all levels of the stack—from modeling to debugging data pipelines or UX prototypes.
-
Passion for fast-paced, collaborative work and an eagerness to ship working systems—not just write about them.
-
Availability for in-person work in the San Francisco Bay Area.
Why This Role?
You’ll have the chance to:
-
Drive core innovation in one of the most ambitious areas of applied AI.
-
See your research make it into users’ hands—fast.
-
Shape both the culture and direction of a company in its earliest stages.
-
Work with peers who combine academic depth with startup speed.
